CULTURAL STEREOTYPES AND NORM OF RECIPROCITY 11/03/2014 1 Allingham and Sandmo Model Allingham and Sandmo (1972): Tax evasion is a product of the risk of detection and punishment in the form of tax penalties Empirical data challenge this purely economic model: 1. Higher level of income reporting than the expected utility model would predict. 2. “Milano-Palermo puzzle”: compliance in cities, like Milan and Palermo, under the same fiscal and enforcement regime should be similar, but findings disconfirm this prediction 2 “Milano-Palermo puzzle” • • Since Milan and Palermo are under the same fiscal and enforcement regime, the Allingham and Sandmo model predicts a similar level of compliance. BUT this prediction is disconfirmed by empirical evidence. 10% Vs 66% Tax evasion 3 WHY? The geographical dimension of tax evasion has received different explanations: – tax evasion is a compensating behavior related to the level of services provided; – manifestation of inadequate civic concern (Putnam et al. 1993); – reciprocity to other taxpayers : citizens choose to contribute to the common good by paying their taxes, if they think (or they know) that others are doing the same (Fishbacher et al. 2001). 4 How to test these predictions? The behavioral diversity among Italian regions concerning tax behavior may be due to 1) regional differences in the perception of the efficiency of the state; 2) regional differences in terms of morality and values; 3) effect of cultural stereotypes and information about the behavior of other tax payers (norm of reciprocity) on tax conduct. Design: Good and Bad Institutions SVO New design: cultural stereotypes and norm of reciprocity 5 The “cultural stereotypes and norm of reciprocity” design is aimed to test: • • • • PRESENCE AND EFFECT OF CULTURAL STEREOTYPES ON TAX PAYERS BEHAVIOR: Test if any difference in terms of tax conduct among cities set in different Italian regions is perceived by subjects (i.e., perception of cultural stereotypes) and how this perception affects tax payers’ decisions. EFFECT OF INFORMATION ABOUT THE CONDUCT OF OTHERS ON TAX PAYERS BEHAVIOR (norm of reciprocity): Test if receiving information about the behavior of other tax payers has any effect on subjects’ willingness to pay. EFFECT OF INTRODUCING FISCAL FEDERALISM: Test the effect on tax payers’ behavior when introducing fiscal federalism in Italy, i.e., taxes are collected at a regional basis and money spent where collected. PRESENCE OF REGIONAL DIFFERENCE IN VALUES AND ATTITUDES TOWARDS OTHERS and how these may affect subjects’ willingness to pay taxes. 6 20 subjects Milan Rome 20 subjects Due to the complexity of the experiment, the program will not be in Ztree 7 Milan Salerno 8 Rome Salerno 9 Experiment: 10 stages • Clerical Task • Stages 1-3: COMMON POT (ITALIAN ACTUAL SITUATION) • Clerical Task • Stages 4-7: SEPARATE POT (FISCAL FEDERALISM) • Stage 8: Risk task • Stage 9: SVO • Stage 10: Result • Stage 11: Questionnaire 10 Clerical task • Subjects will perform the first clerical at the very beginning of the experiment • 2 minutes (practice round) • 7.5 minutes (real clerical task) 11 Stage 1 • Round1: – Information about auditing and punishment (5% pun probability) – COMMON FUND (i.e., the 2 groups are informed that money collected through taxation will be put in a general fund, doubled and redistributed to all participant equally ) • • Round2: subjects are asked to report a certain amount of their income - 0% and 100% - and a tax rate of 30%. Round3: elicitation of expectations about separate compliance: expectations on the behavior of players belonging to group x and y separately (incentivized) The 2 groups are informed that money collected through taxation will be put in a general fund, doubled and redistributed to all participant equally How much do you expect members of your own group and members of the other group declared on average in this stage?” 0%; 10-30%; 30-60%; 60-90%; 100% 12 Stage 1 How: AIMED TO TEST: PRESENCE OF CULTURAL STEREOTYPES: EFFECT OF CULTURAL STEREOTYPES ON SUBJECTS’ BEHAVIOR: IF RESULTS ARE DUE TO CULTURAL STEREOTYPES or SIMPLY TO GROUPNESS? Elicitation of expectations about separate compliance (i.e., compliance of group x and group y). Comparing the results of Stage1 with another treatment S1 with 2 group of subjects belonging to the same region/city Comparing the results of S1with another treatment S2 with only 1 group of subjects from the same region 13 Stage 2 • Round1: – Information about auditing and punishment (5% pun probability) – COMMON FUND – Information about common compliance (i.e., not divided by groups) • • Round 2: subjects are asked to report a certain amount of their income - 0% and 100% - tax rate of 30% Round3: elicitation of expectations about separate compliance: expectations on the behavior of players belonging to group x and to group y (divided by group)(incentive for the accuracy of the answer and feedback about its accuracy at the end of experiment) Subjects are provided with the following information: in stage 1, the average level of compliance of the 2 groups has been … How much do you expect members of your own group declared on average in this stage? 0%; 10-30%; 30-60%; 60-90%; 100% …and members of the other group? 14 Stage 2 AIMED TO TEST: EFFECT OF INFO ABOUT COMMON COMPLIANCE ON SUBJECTS’ EXPECTATIONS AND BEHAVIOR: POSSIBLE OUTCOME: The two groups influence each other and converge to a common behavior (conditional compliance): a decrease of compliance in the group which believes to be more compliant (negative effect) and/or an increase of compliance in the group which believes to be less compliant (positive example) (norm of reciprocity in place). HOW: Comparing subjects’ choice and their expectations in S1with those of S2 15 Stage 3 • Round1: – Information about auditing and punishment (5% pun probability) – COMMON FUND – Information about separate compliance (i.e., divided by group) • • Round2: subjects are asked to report a certain amount of their income - 0% and 100% - tax rate of 30%. Round3: elicitation of expectations about separate compliance: expectations on the behavior of players belonging to group x and to group y (divided by group)(incentivized) Subjects are provided with the following information: in stage 2, the average level of compliance of your own group and of the other group has been … How much do you expect members of your own group declared on average in this stage? 0%; 10-30%; 30-60%; 60-90%; 100% …and members of the other group? 16 Stage 3 AIMED TO TEST: EFFECT OF INFO ABOUT SEPARATE COMPLIANCE ON BELIEFS AND BEHAVIOR POSSIBLE OUTCOME: Knowing that its own group level of compliance is higher/lower than that of the other group may lead either to HOW: Comparing subjects’ choice and their expectations in S2with those of S3 EFFECT ON THE CHOICE: 1) GROUPS CONVERGENCE (norm of reciprocity in place). 2) GROUPS POLARIZATION and EMERGENCE OF GROUP IDENTITY. EFFECT ON EXPECTATIONS: 1) ADAPTIVE HOPES AND FEARS: confirmed empirical expectations will lead one to hold more EXTREME empirical expectations in the future: i.e., subjects express through their expectations their hope and fear about the behavior of the two groups 17 Clerical task • Subjects will perform the clerical task again • 7.5 minutes 18 Stage 4 • Round1: – Information about auditing and punishment (5% pun probability) – SEPARATE FUNDS – NO INFO • • Round2: subjects are asked to report a certain amount of their income 0% and 100% - tax rate of 30%. Round3: elicitation of expectations about the level of compliance of its own group only (incentivized) The 2 groups are informed that money collected through taxation will be put in separate funds, doubled and redistributed to all participant of the group equally How much do you expect members of your own group declared on average in this stage?” 0%; 10-30%; 30-60%; 60-90%; 100% 19 Stage 4 AIMED TO TEST: EFFECT OF INTRODUCING FISCAL FEDERALISM ON THE BEHAVIOR AND EXPECTATIONS OF THE 2 GROUPS POSSIBLE OUTCOMES: 1. Compliance increases in both groups since the general fund is not shared with ‘strangers’ any more 2. Compliance increases even more in the group with the higher level of compliance since the fund has not to be shared with evaders anymore (crowding in effect) 3. Compliance decreases even more in the group with the lowest level of compliance since they know that there is not room for improvement (crowding out effect HOW: SEPARATE POT 20 Stage 5 • Round1: – Information about auditing and punishment (5% pun probability) – SEPARATE FUNDS – Information about the level of compliance of its own group only • • Round2: subjects are asked to report a certain amount of their income 0% and 100% - tax rate of 30%. Round3: elicitation of expectations about the level of compliance of its own group only(incentivized) Subjects are provided with the following information: in stage 4, the average level of compliance of your own group has been … How much do you expect members of your own group declared on average in this stage?” 0%; 10-30%; 30-60%; 60-90%; 100% 21 Stage 5 AIMED TO TEST: EFFECT OF RECEIVING INFORMATION ABOUT THE BEHAVIOR OF OTHER TAX PAYERS IN HIS OWN GROUP ON SUBJECTS’ WILLINGNESS TO PAY POSSIBLE OUTCOMES: HOW: Information about its own group level of compliance only – Reciprocity to other taxpayers : subjects choose to contribute to the general fund by paying their taxes, if they think that others are doing the same, otherwise they will evade. 22 Stage 6 • • • Round1: – Information about auditing and punishment (5% pun probability) – Information about separate compliance – SEPARATE FUNDS (depending on the oucome of the voting) Round2: subjects are asked to report a certain amount of their income 0% and 100% - tax rate of 30%. Round3: elicitation of expectations about expectations on the behavior of players belonging to group x and to group y (divided by group)(incentivized) In stage 5, the average level of compliance of your own group and of the other group has been … How much do you expect members of your own group declared on average in this stage? 0%; 10-30%; 30-60%; 60-90%; 100% …and members of the other group? 23 Stage 6 AIMED TO TEST: WHETHER RECEIVING INFORMATION ABOUT THE BEHAVIOR OF TAX PAYERS IN THE OTHER GROUP AFFECTS SUBJECTS’ WILLINGNESS TO PAY, EVEN WHEN THEY ARE NOT TOGETHER ANYMORE POSSIBLE OUTCOMES: • virtuous people are tempted to behave in a less virtuous way if they know that people in other regions are not so honest • virtuous groups keep their good behavior because they are proud of themselves. HOW: Info about the 2 groups separaterly 24 Stage 7 • Round1: – Information about auditing and punishment (5% pun probability) – Information about separate compliance – VOTING FOR RATIFYING OR NOT THE SEPARATE POT – SEPARATE/COMMON POT (depending on the oucome of the voting) • In stage 6, the average level of compliance of your own group and of the other group has been … Round2: subjects are asked to report a certain amount of their income - 0% and 100% - tax rate of 30%. 25 Stage 7 AIMED TO TEST: PREFERENCE FOR FISCAL FEDERALISM: • How does the behavior of people in the other group affect the choice of ratifying or nor fiscal federalism? • Which regions prefer fiscal federalism? HOW: • Info about the 2 groups separaterly • Result of voting about mantaining or not fiscal federalism POSSIBLE OUTCOMES: – 1. Only rich regions want fiscal federalism, – 2. also the poor ones 26 Stages 8-11 • SVO: The same as in the baseline experiment • Lottery • Results from all of the tasks are shown: – AUDITING AND PUNISHMENT: Subjects will be informed about the auditing and punishment received – EXPECTATIONS: Subjects will receive feedbacks about their answers and paid for their accuracy • Questionnaire 27 “Cultural stereotypes and norm of reciprocity” design The behavioral diversity among Italian regions concerning tax behavior may be due to 1) regional differences in the perception of the efficiency of the state; 2) regional differences in terms of morality and values; 3) effect of cultural stereotypes and information about the behavior of people living in different regions on tax conduct. Design: Good and Bad Institutions SVO New design1 28